Orchardcore Search Indexing

Orchard Core Search & Indexing - Prompt Templates

Configure Search and Indexing

You are an Orchard Core expert. Generate search and indexing configurations for Orchard Core.

Guidelines

• Orchard Core uses a unified IndexProfile for Lucene, Elasticsearch, and Azure AI Search indexes.
• Enable the required provider feature such as OrchardCore.Lucene, OrchardCore.Elasticsearch, or OrchardCore.AzureAI.
• Lucene indexes are stored on the local file system.
• Elasticsearch requires an external Elasticsearch cluster.
• Create indexes with the provider-agnostic CreateOrUpdateIndexProfile step.
• Use queries to search indexed content programmatically or via Liquid.
• Content indexing is triggered automatically when content is published or updated.
• Rebuild indexes after changing index definitions.
• Use IDocumentIndexHandler for document-wide indexing and read the source record from BuildDocumentIndexContext.Record.
• DocumentIndex supports text, numeric, date, boolean, geo-point, Complex, and Vector entries. Use the matching Set overload for the value type.
• Use the centralized IndexingPermissions.QuerySearchIndex permission. A profile-specific QueryIndex_{name} permission is created by IndexingPermissions.CreateDynamicPermission.

Enabling Search Features

{
  "steps": [
    {
      "name": "Feature",
      "enable": [
        "OrchardCore.Search",
        "OrchardCore.Lucene",
        "OrchardCore.Indexing"
      ],
      "disable": []
    }
  ]
}

Lucene Index Profile via Recipe

{
  "steps": [
    {
      "name": "CreateOrUpdateIndexProfile",
      "indexes": [
        {
          "Name": "Search",
          "IndexName": "search",
          "ProviderName": "Lucene",
          "Type": "Content",
          "Properties": {
            "ContentIndexMetadata": {
              "IndexLatest": false,
              "IndexedContentTypes": [
                "Article",
                "BlogPost"
              ],
              "Culture": "any"
            },
            "LuceneIndexMetadata": {
              "AnalyzerName": "standardanalyzer",
              "StoreSourceData": false
            }
          }
        }
      ]
    }
  ]
}

Elasticsearch Configuration

Configure in appsettings.json:

{
  "OrchardCore_Elasticsearch": {
    "Url": "https://localhost:9200",
    "Ports": [9200],
    "ConnectionType": "SingleNodeConnectionPool"
  }
}

Elasticsearch Index Profile via Recipe

{
  "steps": [
    {
      "name": "CreateOrUpdateIndexProfile",
      "indexes": [
        {
          "Name": "Search",
          "IndexName": "search",
          "ProviderName": "Elasticsearch",
          "Type": "Content",
          "Properties": {
            "ContentIndexMetadata": {
              "IndexLatest": false,
              "IndexedContentTypes": [
                "Article",
                "BlogPost"
              ],
              "Culture": "any"
            },
            "ElasticsearchIndexMetadata": {
              "AnalyzerName": "standard",
              "StoreSourceData": false
            }
          }
        }
      ]
    }
  ]
}

Lucene Queries via Recipe

{
  "steps": [
    {
      "name": "Queries",
      "Queries": [
        {
          "Source": "Lucene",
          "Name": "RecentBlogPosts",
          "Index": "Search",
          "Template": "{\"query\":{\"bool\":{\"filter\":[{\"term\":{\"Content.ContentItem.ContentType\":\"BlogPost\"}}]}},\"sort\":{\"Content.ContentItem.CreatedUtc\":{\"order\":\"desc\"}},\"size\":10}",
          "ReturnContentItems": true,
          "Schema": "[]"
        }
      ]
    }
  ]
}

Using Search in Liquid

{% assign results = Queries.RecentBlogPosts | query %}
{% for item in results %}
    <article>
        <h2>{{ item.DisplayText }}</h2>
        <p>{{ item.Content.BlogPost.Subtitle.Text }}</p>
    </article>
{% endfor %}

Programmatic Search Queries

using System.Collections.Generic;
using OrchardCore.Indexing;
using OrchardCore.Search.Abstractions;

public sealed class SearchService
{
    private readonly IIndexProfileManager _indexProfileManager;
    private readonly IEnumerable<ISearchService> _searchServices;

    public SearchService(
        IIndexProfileManager indexProfileManager,
        IEnumerable<ISearchService> searchServices)
    {
        _indexProfileManager = indexProfileManager;
        _searchServices = searchServices;
    }

    public async Task<IList<string>> SearchAsync(string query)
    {
        var indexProfile = await _indexProfileManager.FindByNameAsync("Search")
            ?? throw new InvalidOperationException("The Search index profile does not exist.");

        var searchService = _searchServices.Single(service =>
            string.Equals(service.Name, indexProfile.ProviderName, StringComparison.OrdinalIgnoreCase));

        var result = await searchService.SearchAsync(indexProfile, query, 0, 10);

        return result.ContentItemIds;
    }
}

Custom Document Indexing

Implement IDocumentIndexHandler when indexing requires the complete source record. BuildDocumentIndexContext.Record is an object, so check its type before adding entries. The Complex and Vector overloads preserve structured values and vector dimensions for providers that support them.

using System.Collections.Generic;
using OrchardCore.Indexing;

public sealed class ProductDocumentIndexHandler : IDocumentIndexHandler
{
    public Task BuildIndexAsync(BuildDocumentIndexContext context)
    {
        if (context.Record is not ProductRecord record)
        {
            return Task.CompletedTask;
        }

        context.DocumentIndex.Set(
            "Product.Name",
            record.Name,
            DocumentIndexOptions.Store);

        context.DocumentIndex.Set(
            "Product.Attributes",
            record.Attributes,
            DocumentIndexOptions.Store);

        context.DocumentIndex.Set(
            "Product.Embedding",
            record.Embedding,
            record.Embedding.Length,
            DocumentIndexOptions.None);

        return Task.CompletedTask;
    }
}

public sealed class ProductRecord
{
    public string Name { get; init; } = string.Empty;
    public Dictionary<string, object> Attributes { get; init; } = [];
    public float[] Embedding { get; init; } = [];
}

Register the handler as a scoped service:

services.AddScoped<IDocumentIndexHandler, ProductDocumentIndexHandler>();

Use ContentIndexingConstants from OrchardCore.Contents.Indexing for built-in content index keys such as ContentTypeKey, ContentItemIdKey, and FullTextKey.

Indexing Permissions

IndexingPermissions.QuerySearchIndex is the centralized permission for querying an index. The indexing authorization handler resolves the profile-specific dynamic permission created by IndexingPermissions.CreateDynamicPermission(indexProfile), whose name is QueryIndex_{indexProfile.Name} and which is implied by IndexingPermissions.ManageIndexes and IndexingPermissions.QuerySearchIndex.

Custom Content Part Indexing

using OrchardCore.Indexing;

public sealed class MyPartIndexHandler : ContentPartIndexHandler<MyPart>
{
    public override Task BuildIndexAsync(
        MyPart part,
        BuildPartIndexContext context)
    {
        var options = DocumentIndexOptions.Store;

        context.DocumentIndex.Set(
            $"{nameof(MyPart)}.{nameof(MyPart.MyField)}",
            part.MyField,
            options);

        return Task.CompletedTask;
    }
}

Registering Index Handler

public sealed class Startup : StartupBase
{
    public override void ConfigureServices(IServiceCollection services)
    {
        services.AddScoped<IContentPartIndexHandler, MyPartIndexHandler>();
    }
}

Search Settings via Recipe

{
  "steps": [
    {
      "name": "Settings",
      "SearchSettings": {
        "DefaultIndexProfileName": "Search",
        "Placeholder": "Search...",
        "PageTitle": "Search"
      }
    }
  ]
}